The long term objectives are to develop strategies for early prediction of auto-immune diabetes in the general population, to integrate these predictive strategies with prevention trials, and to use genetic and prospective immunologic data to better understand the pathophysiology. The disease has different phenotypes in sporadic compared to familial cases. Based on family studies, it is assumed that auto-antibodies persistent and that most subjects with high risk marker combinations do progress to IDDM. However, the relevance of family studies is weakened since subjects have greater genetic risk and since only the later stages of pre-diabetes are observed. About 85% of new cases are sporadic, with the highest incidence in the second decade of life. Environmental and non- HLA genetic factors modify diabetes risk and influence non-progression of islet auto-immunity, age of IDDM onset, and immunologic manifestations of the disease. Therefore, we propose that a detailed prospective study of teenagers without familial IDDM is necessary, and requires a large population due to low prevalence. The Washington State Diabetes Prediction Study has already screened 3992 teenage school children, and followed those with risk markers in detail. Three developed IDDM; the rest remain normoglycemic, some despite multiple auto-antibodies and high risk HLA DQ alleles. Since the initial screen, new protein auto-antigens have been cloned, and novel human genome markers reveal non-HLA candidate disease loci. We propose to complete auto-antibody and genetic screening on the entire cohort. Those persistently marker-positive will continue to be followed by beta cell function to identify progressors and non- progressors. They will have detailed immunogenetic testing including GAD65, insulin, and ICA512 auto-antibody levels, subclasses and epitopes, T-cell proliferation and cytokine production, and genotyping at all IDDM loci proven to be associated with disease. The subjects were first fled 5 years ago, so the completed study will have a 10-year follow up. Our hypothesis are that genetic screening followed by periodic auto- antibody tests detects progressors more efficiently than auto-antibodies alone, that detailed immunogenetic analyses reveal protection or progression earlier than tests of insulin secretion, and that subjects with beta-cell specific immune responses but not progressing to diabetes represent intermediate phenotypes associated with protective genotypes at susceptibility loci. Crucial interactions with Projects 1 and 2 will integrate detailed HLA-DQ and B-lymphocyte biology and its genetic control into these phenotype-genotype analyses. Project 3 will provide distinct IDDM phenotypes for the same genetic analyses, to contrast with our results in Project 4. Core B will provide large-scale DNA extraction and genotyping support critical to this project. Results will fill critical gaps in knowledge of early events and factors dictating progression or protection in sporadic auto-immune diabetes.